Motility Data Acquisition and Analysis.

Once the flow cell is loaded on the microscope, beads are trapped by a 1,064-nm laser and calibrated for trap stiffness and position within our detector zone. Tethered beads are centered over the complex and subsequently displaced by moving the piezo stage. In the instances where the tether shows evidence of folding and microstructure formation, we continue to move the stage as the tether undergoes extensions and mechanical relaxation before we record biosynthesis rates. As cellulose is synthesized, bead position is recorded at 5 kHz for as long as 10 min. Fiducial 0.75-μm polystyrene beads bound to the coverslip facilitate drift tracking and correction through a custom cross-correlation video tracking algorithm similar to Brady et al. (36). Motility traces are corrected for drift and then decimated to 100 Hz. The actual tether length is calculated with a correction factor of 1/sin(θ), where θ is the incidence angle (angle from a line perpendicular to the coverslip surface) (50). Custom MATLAB scripts determine a velocity and average force for the entire trace. Velocities of less than 0.01 nm s⁻¹ were considered an absence of synthesis. For the force-velocity curve, velocities were binned by force and averaged every 2 pN. Each bin was weighted corresponding to the number of items in each bin for the fit. We fit the weighted, averaged velocities to the general Boltzmann relationship (28). Abrupt extensions or retractions were located by a sliding step-finding MATLAB script that distinguished changes in mean position greater than two SDs from the previous segment’s mean (78). The two sliding segments were 0.2 s in size. Typical force ranges for all control experiments were 3 to 8 pN, within the active range unaffected by force. The randomness parameter was calculated individually for each trace from the variance from the mean trajectory and averaged to find the mean randomness parameter for each given characteristic distance.